AccelerateBench v1.1 - Geometric Mean Speedup: leaderboard
Geometric mean speedup (x, with incorrect or unaccelerated runs assigned 1.0x) across the 32 HPC program optimization tasks of AccelerateBench v1.1 (arXiv:2611.20882v1, Table 1).
Metric: Geometric mean speedup (x). Source: arxiv.org. Saturation forecast: Around February 2028. 16 models tracked. Updated 21 Nov 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Claude Opus 5.5 | High | 3.94 | #9 |
| 2 | Claude Sonnet 5.5 | High | 3.83 | #24 |
| 3 | Kimi K3 | High | 2.47 | #61 |
| 4 | GPT-6.1 Sol | High | 2.42 | #15 |
| 5 | hy4 | High | 2.41 | #40 |
| 6 | GPT-6 Astra | High | 2.4 | #10 |
| 7 | Hy4 preview | High | 2.31 | #74 |
| 8 | DeepSeek V4.1 Flash (1017) | High | 2.28 | #47 |
| 9 | GLM-5.3 Flash | High | 2.2 | #133 |
| 10 | GLM-5.3 | High | 2.2 | #72 |
| 11 | DeepSeek V4 Pro | High | 2.19 | #196 |
| 12 | Grok 4.6 | High | 2.13 | #50 |
| 13 | GLM-5.2 | High | 1.98 | #150 |
| 14 | DeepSeek V4 Flash | High | 1.92 | #267 |
| 15 | GPT-6.1 Terra | High | 1.92 | #25 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=acceleratebench-v1-1-geometric-mean-speedup · How It Works · Data refreshed daily, snapshot 2026-11-21.